Cloud Applications Drive Advanced Networking
Enterprises must rethink network management in the cloud computing world. This new reality is driven by the rise of software defined networking, the virtualization of everything and a business imperative to create and deploy even newer information delivery models. With the entire infrastructure stack now fully programmable, the need to integrate the network with cloud compute services is essential. Another important environmental change is that both servers and applications have joined data with respect to mobility. This leads to an increased need to categorize those assets and apply appropriate security controls with respect to business/mission functions, data user role, location of access, legal/regulatory guidelines and user devices. In meeting the challenges of mobility and globalization, enterprises must address their lack control or complete visibility into the networks they use.
Organizations often lose the economic and operational benefits of the cloud if the network provider lacks the agility afforded by network management automation. Requirements to support cloud brokerage and to bridge services between different cloud service vendor marketplaces are also coming to the forefront. Network performance expectations that are driving more advanced network management requirements include:
- Data-intensive analytics;
- Parallel and clustered processing;
- Remote expert consultation and collaboration; and
- Community cloud services.
Real-time network analytics is also needed to manage user application experience. Service capabilities should include:
- Network latency management – Acceptable response times on the web have steadily been decreasing. While 10 seconds used to be considered reasonable, a 2009 study showed that 40 percent of visitors will abandon a web page if it takes more than three seconds to load. Interactions that take 1 second or less maintain a user’s seamless flow of thought.
- Network Reliability – Reliability is an attribute of any computer-related component that consistently performs according to its specifications. It has long been considered one of three related attributes that must be considered when making, buying, or using a computer product or component.
- Network Performance – Metrics of the service quality of a network as seen by the customer. These metrics usually address bandwidth, throughput, latency, jitter and error rate. All of these factors, coupled with user requirements and user perceptions, play a role in determining the perceived utility of a network connection.
- Internet path visibility – Cloud computing’s placement of critical infrastructure components outside traditional network boundaries increases enterprise IT dependence on the complex interactions between enterprise and public IP networks. In order to ensure reliable application delivery, this fact drives a requirement to provide visibility into the routing and traffic dynamics spanning enterprise and Internet domains to the enterprise network manager.
- Automated routing ruleset – the real-time use of genetic adaptive routing algorithms (e.g. path mutation and path crossover) to identify and make available new network paths for message propagation
- Dynamic bandwidth allocation – Dynamic bandwidth allocation is a technique by which traffic bandwidth in a shared telecommunications medium can be allocated on demand and fairly between different users of that bandwidth. This is a form of bandwidth management, and is essentially the same thing as statistical multiplexing.
These capabilities must be broadened in order to deal with dynamic operational or application demands of even the smallest organization. Cloud-based service-level agreement will eventually require the network service provider to enable both planned and unplanned bandwidth variations as well. Service-level agreement will also be needed for in-sequence packet delivery and jitter. In the near future, these types of network quality of service metrics will be crucial to enterprises when they consider moving to new network operators.
New identity access and authorization use cases are also challenging current network design best practices. Virtual equivalences to physical and infrastructure-centric protection norms are mostly non-existent. An inability to enforce current, geographic-centric legal frameworks in the virtual world poses a major barrier to the development of a practical, economical and international cloud computing legal framework. Some of the operational issues dependent on major advances in this realm include:
- Mobile device-based access to everything;
- Mobile office productivity;
- The broadening project economy;
- Virtual organizations; and
- Virtual reality.
Network management foundational changes include the discovery of new network traffic patterns. The client-server past is quickly morphing into a future in which server-to-server traffic predominates. This transition unveils location-independent endpoints on both sides of a service or transaction. The growing operational use of content delivery network “walled gardens” for network quality-of-service management contributes to these changing patterns.
When accepting this network management leap forward, enterprises need vendor partners that provide tools necessary to manipulate the operational levers of cloud networking. Your network may, in fact, be impeding your efforts to embrace cloud, big data, mobile, social and security technology trends. Primary among these tools is a software-defined networking (SDN) solution. Such a solution can virtualize your network and create centralized controls to provide more agile and flexible network services. When built on open standards, a software-defined network helps you:
- Realize faster time to market by deploying networks in minutes
- Reduce management complexity and operating expenses through automated controls
- Enhance security by inserting patterns in the software
- Boost innovation by using cloud resources more easily
- Build an agile platform to enable analytics and gain actionable business insights
Dynamic network optimization is also a key network operational lever. A key technology in this area is network function virtualization. In addition, network architecture standardization can be used to improve availability, temper costs and reduce the risk of security breaches